The Natural History of Portal Venous System Aneurysms

Objective: Portal venous system aneurysms (PVA) are increasingly diagnosed on axial imaging. However, the natural history of these aneurysms is poorly understood, and reports are limited to case series.

Methods: A search for PVA in radiology reports was performed between 01/2010 and 05/2022 followed by manual review. Portal vein aneurysms were defined as an anterior-posterior diameter greater than 1.5 cm and 1.9 cm in patients without and with cirrhosis, respectively. Aneurysm growth was defined as greater than twenty percent of the original size. Patient demographics, comorbid conditions, and PVA outcomes were abstracted.

Results: Sixty-two aneurysms were identified in 59 patients, and involved the portal vein (n=30, 48.4%), splenic vein (n=17, 27.4%), superior mesenteric vein (n=4, 6.5%), and portal confluence (n=11, 17.7%). While 23 (37.1%) aneurysms were idiopathic, the remaining 39 (62.9%) were associated with portal hypertension (n=25, 40.3%), pancreatitis (n=4, 6.5%), and prior liver transplant (n=5, 8.1%) (Table 1). The initial median size was 2.6 cm (1.1–4.7 cm) with larger aneurysms at the confluence (p=0.01). Of the 39 PVA with follow up over 4.75 years (4 months-16.6 years), the median growth was 0.2 cm (-2.6–2.4 cm). Thirteen (33.3%) aneurysms grew more than 20% during a 7.5-year follow up (8 months-16.6 years). Growing PVA were associated with portal hypertension (84.6%, p=0.003), current tobacco use (38.5%, p=0.02), and thrombosis during follow-up (46.2%, p=0.04). Twelve (19.4%) thrombosed, predominantly in patients with cirrhosis (n=9, 75.0%). Of the 10 thrombosed PVA with follow up over 3.8 years (9 months-15.4 years), the median growth was 0.8 cm (-0.7–1.9 cm). Complications of PVA thrombosis included abdominal pain (n=2, 16.7%), intestinal ischemia (n=1, 8.3%), and variceal bleeding (n=3, 25.0%). Four (33.3%) patients were treated with anticoagulation. No aneurysms ruptured. While most patients were managed conservatively, 7 underwent intervention where 6 (85.7%) were symptomatic with a median diameter of 3.7 cm (Table 2). Intervention included resection (n=4, 6.5%), ligation (n=1, 1.6%), and coil embolization (n=2, 3.2%). There was one case of aneurysm recurrence 20 years following resection and one case of post-operative mortality.

Conclusions: Most PVA only require surveillance as two-thirds remain stable. Nearly 20% of PVA thrombosed but none ruptured. As the majority thrombosed without issue, treatment with anticoagulation should be based on symptoms and extent of thrombosis.